Mine Google Play store reviews

[megakevin]

Let's start mining the Google Play store's user comments/reviews/ratings. We can use this to find correlations between the attack surface and certain characteristics that we identify from this data.

Through natural language processing or topic modeling we can determine what the users are talking about in these comments and identify instances that are security or quality related.

Then we can see what the attack surface can tell us about that or what that can tell us about the attack surface. If there is a relationship there.

[megakevin]

Wrote a python script to download the user reviews of a given apk: https://github.com/megakevin/android-attack-surface/blob/master/googleplay-scraper/get_reviews.py

Works well but google doesn't like the automated requests to their servers and thus blocks from time to time.

This script will be called by a main bootstrapper script after the main scraper has put the apks information into the database.

Still need to:

• Implement a CLI that allows it to receive an apk ID as input.
• Write the downloaded info into a database instead of text files as it does right now.
• Implement a delay between requests so that google wont block us.

[megakevin]

https://github.com/megakevin/android-attack-surface/blob/master/googleplay-scraper/get_reviews.py

[megakevin]

The code that does this is written. Only need to run it.